Parham Zendehdel Nobari

PhD Student · ETH Zürich

Working on Applied Computer Vision 

I am a PhD student at the Chair of Infrastructure Management at ETH Zürich, led by Prof. Bryan T. Adey, where I work on the Horizon Europe project SAFE-MOVE. My research uses computer vision and remote sensing for road safety: I combine vehicle sensor data with imagery to find the built-environment features and near-misses that put cyclists, pedestrians and micromobility users at risk.

I hold an MSc in Civil Engineering from ETH Zürich. For my thesis, supervised by Prof. Catherine De Wolf at the Chair of Circular Engineering for Architecture, I trained a computer vision model on drone imagery to map rooftop materials across more than 1,200 buildings in Geneva. My work spans end-to-end deep learning pipelines, from data acquisition and annotation to training and deployment, as well as aerial photogrammetry, point cloud processing and fusing imagery with geospatial data.

Before my PhD, I worked in industry on BIM-based project delivery and building-technology innovation. Earlier, I studied civil engineering at Amirkabir University of Technology.

Portrait of Parham Zendehdel Nobari

Publications


Data Science for Transportation (Springer) cover illustration

Exploring Deep Learning Approaches for Short-Term Passenger Demand Prediction

Zahra Ghandeharioun, Parham Zendehdel Nobari, Wenhui Wu

Data Science for Transportation Springer Nature, vol. 5, 2023

Detritus: Multidisciplinary Journal for Waste Resources & Residues

Evaluation of the application of municipal solid waste incinerator (MSWI) ash in civil engineering using a sustainability approach

Setareh Seraj, Morteza Nikravan, Ali A. Ramezanianpour, Parham Zendehdel

Detritus vol. 9, 2020

Selected Projects


Rooftops segmented by material (metal sheets, clay tiles, gravel, bituminous sheets) from drone imagery

Urban-Scale Rooftop Material Detection with Computer Vision

Master Thesis ETH Zürich, D-BAUG, IBI, 2023

End-to-end pipeline on UAV aerial photogrammetry imagery training a Mask R-CNN instance segmentation model (PyTorch, Detectron2) to classify and quantify rooftop materials across 1,200+ buildings in Geneva, producing a georeferenced material cadastre to support circular construction.

Render of Flexilibres, the circular timber redesign of the Equilibres building

Custom BIM Viewer for Material Passports

Master project ETH Zürich, D-BAUG, IBI, 2021

A web BIM viewer on Autodesk Platform Services that puts a building's material passport onto its 3D model. Built for Flexilibres, a circular, low-carbon timber redesign of the Equilibres building, it colors every element by passport metrics such as embodied CO2, lifespan and reuse potential, summarizes them by building layer or material, and lets you filter the model and inspect each element's passport.

Energy Compacts metrics dashboard: world map of coverage regions and SDG 7 metric tiles

Energy Compacts Visualization Tools

Hack4Good ETH Zürich, Analytics Club, 2026

Two browser-based tools for exploring the UN Energy Compacts. A force-directed network graph (D3.js) of 169 compacts, 1,204 actors and 2,067 relationships shows who works with whom, and an interactive dashboard turns compact data from Excel or CSV into SDG 7 metrics, a filterable world map and target-year views. Both run entirely in the browser, so no data leaves the user's machine, and export as standalone web pages.

Education


Professional Experience


Teaching


Selected Achievements